Using the Chou's 5-steps rule to predict splice junctions with interpretable bidirectional long short-term memory networks. (January 2020)

• Record Type: Journal Article
• Authors: Dutta, Aparajita; Dalmia, Aman; R, Athul; Singh, Kusum Kumari; Anand, Ashish
• Is Part Of: Computers in biology and medicine. Volume 116(2020)
• Publication Date: 2020-01
• Subjects: Splice junction prediction -- Visualization -- Bidirectional long short-term memory networks -- Attention -- Occlusion -- Omission -- Integrated gradients -- Smooth gradients; Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2019.103558
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: Neural models have been able to obtain state-of-the-art performances on several genome sequence-based prediction tasks. Such models take only nucleotide sequences as input and learn relevant features on their own. However, extracting the interpretable motifs from the model remains a challenge. This work explores various existing visualization techniques in their ability to infer relevant sequence information learnt by a recurrent neural network (RNN) on the task of splice junction identification. The visualization techniques have been modulated to suit the genome sequences as input. The visualizations inspect genomic regions at the level of a single nucleotide as well as a span of consecutive nucleotides. This inspection is performed based on the modification of input sequences (perturbation based) or the embedding space (back-propagation based). We infer features pertaining to both canonical and non-canonical splicing from a single neural model. Results indicate that the visualization techniques produce comparable performances for branchpoint detection. However, in the case of canonical donor and acceptor junction motifs, perturbation based visualizations perform better than back-propagation based visualizations, and vice-versa for non-canonical motifs. The source code of our stand-alone SpliceVisuL tool is available at https://github.com/aaiitggrp/SpliceVisuL . Graphical abstract: Highlights: We employ BLSTM network with attention for the prediction of splice junctions. The proposed architecture, named SpliceVisuL, achieves state-of-the-art performance. Some visualization techniques are redesigned to comprehend genome sequences. Features learnt by the model are extracted and validated with the existing knowledge. A comparative study of the visualizations is done in terms of the learnt features.